In 2000, George W. Bush campaigned as a "uniter, not a divider." It didn't pan out that way. In four years time, the electorate was evenly divided, with about half the country favoring his reelection and the other half opposing it.

This year, President-elect Obama campaigned on moving the country past its political divisions to focus on what unites it. The results from this month's election suggest he might have his work cut out for himself on this front. While his popular vote and Electoral College victories were decisive, there are indications that the electoral polarization we have seen in the Bush years persists.

To begin, we need a way to measure polarization, which is simply the accentuation of differences. So, the greater the differences among factions in the electorate, the more polarized we can say the whole electorate was. We'll put forth two ways to measure this concept.

First, we'll take an unweighted average of Obama's share of the vote from every state plus D.C. We'll use this as a baseline to calculate the standard deviation, which is a measure of the variation around the average. That's what we're really interested in. The greater the standard deviation, the more the states varied around the average, the more accentuated were their differences, and so the more polarization there was.

We replicate this method for every presidential election going back to 1948, which enables us to compare this cycle to the last fifteen:

First, notice how the graph confirms widely-held beliefs about polarization. We see that it was higher in the 1960s than the '50s or '70s. Note 1964 in particular. While Lyndon Johnson won a large nationwide victory, the South swung heavily against him. He won 61.1% of the national popular vote, but just 12.9% in Mississippi. This is polarization.¹

Second, notice that it was up in 2000 and 2004. Between Nixon's reelection and Clinton's reelection, polarization was as low as it ever was in the 20th century. However, it jumped up in 2000, and remained steady in 2004. Again, this squares with common knowledge of the last eight years - which holds that there has been more polarization recently. Now, turn to the final point on the chart, 2008. Polarization was about the same this year as in 2000 or 2004. In fact, by this metric, it actually ticked up a little bit.²

This metric uses an unweighted average of a victorious candidate's share of the vote in all states as a first step in calculating polarization. This means that, regardless of population, each state plus D.C. counts for 1/51 of the victor's average. So, it might be helpful for our second metric to use the nationwide popular vote as a first step.³

We'll do that by taking the victor's share of the nationwide vote, and then by counting the number of states where his share of the statewide vote was at least ten points higher or ten points lower than his nationwide vote. This can tell us how many "polarized" states there were that year. Again, we'll replicate this method going back to 1948, and graph the results:

While there are some minor differences, this chart generally squares with the previous one. We again find polarization peaking in the 1960s - by this measure 1968, not 1964, was the peak year. From the 70s through the 90s, it again settled down, only to rise again with George W. Bush. Once again, we see it go a little higher with Barack Obama.⁴

Let's push this analysis forward by examining which states have been polarized. We'll use our second metric for this. Instead of counting the states, we'll color code maps. The following picture does this for all elections going back to 1976. States shaded blue are pro-Democratic "polarized" states. Those shaded red are pro-Republican.⁵

Typically, strong Republican states can be found in the Mountain Region. Utah is consistently red. Idaho and Wyoming are frequently so. We also see states in the West North Central region tilt red with some frequency: Kansas, Nebraska, and the Dakotas. George W. Bush's elections were especially polarizing because he swung all of these states into his column, rather than just a few. Meanwhile, four states in the northeast swung strongly against him: Massachusetts, New York, Rhode Island, and Vermont.

Many of these states again showed a high degree of polarization in 2008. This time, however, they were joined by most of the remaining states in the South Central regions: Arkansas, Kentucky, Louisiana, Mississippi, and Tennessee. Never before have we seen these states vote so heavily against a victorious Democrat. Ditto West Virginia, which went for Michael Dukakis in 1988.

So, by both metrics, we find that polarization did not decline this cycle. It was about where it was in the 2000 and 2004 cycles, a point that's higher than anything we have seen in the last 30 years.

Before we discuss the implications of this, there are some important caveats to note. To start, any argument about a concept like polarization is only as good as the empirical metrics for it. We have two metrics that are pretty good, but any limitations they have will inevitably limit the conclusions we can draw. For instance, these are not metrics for ideological polarization - so we're not discussing that concept here.

Another caveat. We have operationalized the concept of polarization such that we're measuring divergence around some central number, be it the average vote across all 50 states or the national vote. However, we have not yet taken into account the central number itself. It's pretty important. By these metrics, polarization was about where it was in 2000. However, Bush won 47.9% of the vote in 2000 to Obama's 52.7% in 2008. That puts President-elect Obama in a stronger position than his predecessor. President Bush won a polarized election in 2000 with less than half the vote. President-elect Obama won more than half. That makes a big difference.

With these caveats down, what conclusions can we draw from this analysis? First, it is fair to say that this indicates that the political polarization we have seen in recent cycles cannot solely be chalked up to the personality of George W. Bush. Instead, it appears as though there might be a systemic cause, one that accounts for the elections of 2000, 2004, and 2008 displaying polarization regardless of who is on the ballot.

There's a second conclusion to draw. Pundits have been giving a lot of free advice to the President-elect, drawing heavily upon the experiences of exceptional presidents from years gone by. FDR and Lincoln seem to be the most widely referenced. If I were to suggest a previous Chief Executive the new President-elect should study closely, I'd recommend George W. Bush.

The essential job of the President is to be the leader of all the people. He is the only official we all select; accordingly, he's the only one who can claim to represent all of us. That poses a special challenge when the people are polarized, and this has been a problem for the 43rd President. Regardless of one's opinion of George W. Bush, it is fair to say that he has not governed with an eye to those who have strongly opposed him.

When you consider Article II of the Constitution against what past Chief Executives have actually done, it becomes clear that the President's greatest powers are informal, not written down. In many respects, the President does what the people allow him to do. Accordingly, if he begins his tenure with a faction that is disinclined to him, and proceeds to antagonize it - his power can be diminished. That faction can rise in opposition to thwart him, which is easy to do in our system of "checks and balances" that intentionally protects minority rights. I think something like this has happened to Bush over the last few years. What was perhaps a mere disinclination to him in 2000 became much more salient in the subsequent years. This opposition had eroded his informal power by 2005, put the Democrats in control of Congress by 2006, and delivered Obama a sweeping victory by 2008.

Given the data presented here, I think this is a cautionary tale for the President-elect. This does not mean, of course, that he must govern like a Republican from Kentucky. Far from it! It just means he should be aware that there are factions in the country that strongly opposed him, and he should be careful with how he manages these groups. He does not have to do what they want him to do, but he should not overly antagonize them.

The political consequences of that could be harmful. For instance, in the states shaded red in the 2008 picture, there are about 20 white Democratic Representatives and 7 Senators who will stand for reelection at least once in the next eight years. President-elect Obama has to be mindful of them - otherwise, they could go the the way that the New England Republicans have gone in the Bush years.

***
Endnotes

^[1] Prior to 1948, polarization was much higher by either metric. The reason is that the South was a one-party region, voting overwhelmingly Democratic in most every cycle. Occasionally, Republicans who won large nationwide majorities could snag a few Southern states. However, by and large, Republicans ran into the same trouble that Teddy Roosevelt faced in 1904. TR won 56% of the nationwide popular vote that year, but pulled in less than 5% in South Carolina. Conversely, FDR won about the same share of the nationwide popular vote in his four victories while winning 85%-99% of the vote in South Carolina.

^[2] The value for 1964 needs to have a metaphorical asterisk placed next to it because Lyndon Johnson was not on the ballot in Alabama. The standard deviation for that cycle is calculated by taking no value for the state. If he had been on the ballot, he probably would not have done much better than he did in Mississippi, which typically votes in tandem with Alabama. The same goes for 1948 when Harry Truman did not appear on the ballot in Alabama. Including Alabama would increase the standard deviation for both years.

^[3] I also calculated weighted standard deviations (where each state is factored according to its share of the nationwide popular vote) and tracked them over time. The results were quite similar to the unweighted standard deviations:

Again, we see polarization drop in the 70s, 80s, and 90s - only to make a comeback with the elections of George W. Bush and now Barack Obama.

We also see that, overall, polarization is lower by this metric - even though the ratios from year to year are about what they were using the unweighted metric. The reason for this is that a lot of sparsely populated states - Alaska, the Dakotas, D.C., Wyoming, etc. - tend to exhibit polarized behavior in most cycles. In our unweighted average, each of these places counts for 1.9% of the total. Now that we're weighing each state for population - they count for less, typically between 0.2% and 0.3%. This has the effect of pushing the metric down across all years. Of course, because these places always tend to vote the same way cycle after cycle, weighing them does not alter the ratio between cycles, which is what we are really interested in.

^[4] We can tweak our cutoff point, maybe make it +/- 8 points or 12 points, instead of 10 points. That would change the number of polarized states in each cycle, but the crucial point is the same: the last few cycles have seen polarization go up, and in the most recent cycle it is as high as it has been in many decades.

^[5] The picture doesn't capture it, but Washington D.C. would be shaded blue in every election. Alaska would be shaded red in most of them, and Hawaii in many of them. This should condition how we interpret the previous graph. Alaska first voted in a presidential contest in 1960. The District of Columbia voted for the first time in 1964. Since polarization is typical in both places, that essentially inflates the number of polarized states from 1960 to the present by at least one and sometimes three units.

Barack Obama's decisive victory last Tuesday has some wondering whether this was a realigning election.

"Realignment" is an overused term, and some scholars have questioned whether it is a profitable category to apply to elections. Temple University's Robin Kolodny wrote this a few years ago:

Realignment has been in trouble as a theory for explaining party identification and electoral behavior for some time. The most obvious problem is that there has been no full realignment since 1932, and no consensus has emerged on what, if any, partial realignment has taken place in 1968, 1974, 1980, or 1994.

Yale University's David Mayhew wrote a cogent critique of realignment theory in 2004, arguing that the facts don't fit the story so well.

So, let's lower our sights a little bit. Let's put aside the terminology and compare 2008 to three times that, regardless of whether they were realignments, were definitive moments in American electoral history: 1860, 1894-96, and 1932. Realignment or not, it should be profitable to see how today compares to these past times.

*****
1860

Upon James Polk's election in 1844, the Union was equally balanced between slave and free states. The addition of so much territory during his term disrupted that balance. The South wanted to extend slavery to the Pacific. A growing segment in the North wanted to limit it to existing slave states.

The government tried two solutions in four years. The first was the Compromise of 1850. The deal ultimately split the Whig party into regional factions. By 1856, it was gone. The second was the Kansas-Nebraska Act, implemented in 1854. It allowed popular sovereignty to determine whether the Kansas and Nebraska territories would be slave or free, precipitated a violent conflict in Kansas, split the Democrats, and effectively created the Republican Party.

By 1860, the stage was set. The Republicans nominated Abraham Lincoln. Southern Democrats nominated Vice-President John Breckinridge on a pro-slavery platform. Northern Democrats nominated Senator Stephen Douglas on a popular sovereignty platform. Finally, a group of old Whigs and "Know-Nothings" formed the Constitutional Union Party, nominated former House Speaker John Bell, and called for saving the Union.

The following picture shows how this played out. As usual, Republicans are in red and Democrats are in blue. Also, Whigs are in brown, Southern Democrats are in gray, and Constitutional Unionists are in purple:

Lincoln won less than 40% of the popular vote, not having appeared on the ballot in most Southern states, but his Electoral College victory proved how politically powerful a unified North could be: 180 for Lincoln, 72 for Breckinridge, 39 for Bell, and 12 for Douglas.

1894-1896

By the 1880s, the Democrats had returned to electoral competitiveness by accepting many of the political premises of industrial development. The end of Reconstruction and the Panic of 1873 ultimately gave them control over the House for eight of the next ten Congresses. The lone Democratic President of the era - Grover Cleveland of New York - favored the gold standard, which was good for industrial interests in the East but hard on farmers in the South and Midwest.

The grievances of farmers and rural people found expression via the Populist Party (shaded yellow in the subsequent picture), which had become a regional political force by 1892. The economic crisis precipitated by the Panic of 1893 brought these tensions to a head. The midterm election of 1894 saw the GOP pick up 130 House seats, based on big gains in the Northeast and North Central regions.

This was the beginning of a change that would manifest itself on the presidential level in 1896 when William Jennings Bryan captured the Democratic nomination, promising "free silver." His opponent, William McKinley, supported the gold standard. The election of 1896 was fought over the currency issue, and the result produced a sharp industrial-agrarian divide.

Though the South is joined this time by the Mountain West and the Great Plains, the divide again favors the North. McKinley won 271 electors to Bryan's 176.

1932

This was a significant election not simply because the Depression began under the GOP's watch. It also had to do with the party's response. President Herbert Hoover failed to address the crisis to the public's satisfaction. Meanwhile, the Democrats nominated New York Governor Franklin Roosevelt, who had a great last name and a solid reputation of his own, having mobilized his government to fight the Depression in the Empire State.

Unlike in 1860 or 1896, a very broad transregional consensus emerged. As famed newspaper editor William Allen White later observed, the election of 1932 signaled "a firm desire on the part of the American people to use government as an agency for human welfare."

*****

While the particulars of these elections are different, they tell a similar story about the political parties. In all three, the parties had to manage issues of great importance that could not be ignored. This is why we remember Lincoln's "House Divided," Bryan's "Cross of Gold," and Roosevelt's "New Deal." They each took clear stands on issues whose resolutions would determine the course the nation would set.

What's more, there was little room for common ground those years. Either slavery would expand or it wouldn't. Either the government would authorize the free coinage of silver or it wouldn't. Either it would take a more active role in the economy or it wouldn't. Practically speaking, the differences could not be split.

So, these issues upset the normal functioning of the parties. By their nature, parties select issue positions and emphases in pursuit of electoral majorities. Obviously, no party can undertake a full-scale reinvention of itself. However, in pursuit of a majority, it can frequently "finesse" matters. It can slightly alter some positions, it can equivocate or obfuscate on others, and it can emphasize particular issues or personalities depending upon the audience. The goal is to string together an electoral majority among the diverse elements of our large Republic.

In these years, this process was disrupted to some degree. Issues of great salience dominated the political discourse and forced the parties to stake out relatively clear positions. There was little room for finessing. Thus, votes from those years can be seen as opinions on the critical issues more directly than votes from other years.

So, examining the parties and the issues they handled this cycle might help us understand how 2008 stacks up against these three elections. Did the parties behave similarly this year as they did then? Were the issues similar?

I think the answers to both questions are negative, which cuts against the hypothesis that this election was a "realignment." For starters, there was no central, defining issue that disrupted the normal party process. Instead, both candidates covered a variety of issues, few in any depth. There was also a scarcity of clear contrasts between Obama and McCain. Indeed, on the subject that might have emerged as a realigning issue - the financial bailout - they voted the same way.

Relatedly, both candidates made the search for common ground a defining feature of their candidacies. McCain would cite Hillary Clinton just as often as Obama would mention Richard Lugar. There was no House Divided, no Cross of Gold, no New Deal. There was the promise of pragmatic governance and a change in tone toward bipartisan conciliation.

This evidence disfavors the idea that 2008 was like these previous elections. Now, it might be that 2008 was a kind of realignment - perhaps a "partial" one. However, I would return to the above quotation from Robin Kolodny. Adding a qualifier like this strikes me as the sort of inelegance that is tolerated when a theory is losing its explanatory power - like adding epicycles while waiting for the Copernican revolution.

None of this is to claim that the GOP isn't in trouble. For the Republicans, much depends on how well Obama governs. If he governs to the public's satisfaction - the GOP could be in the minority for a while. If he does not - it's return may be speedier.

After making some unfortunate comments to the editorial board of the Pittsburgh Post-Gazette, calling voters in his region "racist" then "redneck," Western Pennsylvania congressman John Murtha finds himself in a tight race. The NRCC and the DCCC have both entered the fray, blanketing the Pittsburgh area with advertising, and Murtha is calling upon Bill and Hillary Clinton to help him carry his district tomorrow.

Murtha's district is Pennsylvania 12th. It covers a vast expanse of Western Pennsylvania, stretching 130 miles from Johnstown in the east to Waynesburg in the southwest corner of the state. His district is the one shaded blue in this picture, courtesy of Sean Oxendine.

As this Post-Gazette article notes, Murtha's district did not always look like this. When he was first elected in 1974, his district, then Pennsylvania's 22nd, looked like this (again courtesy of Sean Oxendine):

Centered in Johnstown, it stretched north instead of west. The shape of the district has changed because Pennsylvania has lost congressional districts over the years - from 27 in the 1970s to 19 in the 2000's.

According to the Post-Gazette, this has contributed to Murtha's problem. Around Johnstown, voters know him well. However, he is not known nearly as well to the voters in the west.

The farther west you go, the deeper the puzzlement at the congressman who called Western Pennsylvania "racist" in an interview with the Post-Gazette editorial board and then apologized by substituting "redneck" in a television interview.

Mr Murtha already had irked veterans and social conservatives in the area by calling for a U.S pullout from Iraq and accusing eight Marines of atrocities in the town of Haditha.

Back home, the "racist" remark is viewed as Mr. Murtha -- who is as famous for his plain speech as for his capacity for handing out federal moolah -- speaking more bluntly than clearly. Yet, in much of the new district, well to the west of Johnstown, there is puzzlement verging on anger.

Murtha's problem is similar to the kind of trouble that many long-serving members of Congress face. When they are first elected to Congress, and for the first few terms, they develop what political scientist Richard Fenno calls a "homestyle." This is a way of interacting with and understanding voters in the district - cultivating good relationships with important constituent groups, appearing at the right events, knowing when to vote with the district and when you can get away with voting against it, and so on.

However, members can face trouble as the district changes. In some instances, the change is induced by new types of voters moving in. In other instances, such as Murtha's, redistricting means new lines and new voters. These new voters do not have the same relationship with the member, which means he must cultivate a bond with them. If he fails to do that, those new voters might ultimately turn against him.

This Post-Gazette article indicates that Murtha has had trouble adapting to his changing district. Voters in Johnstown are more forgiving of Murtha's comments, but to the west his voters are less so because they don't know him as well. This makes intuitive sense. If you've known John Murtha as a man who has brought jobs and federal dollars to the district for 34 years, you have a context for those comments. But if he's new to you, and these comments are some of the first things you've heard from him, they could be very important in shaping your opinion.

So, Murtha's problem this cycle is an extreme example of a common kind of trouble. Congressional districts are not static. They change because of population movements, shifts in the economy, redistricting, and other factors. Members, even long-serving ones, need to adapt to these changes. If they don't, they can face the kind of trouble that Murtha has today.

Will this be enough to end Murtha's career? I don't know. Fortunately for him, he made his bone-headed remark in a year that favors the Democratic Party. That's a lucky break for him - and luck can sometimes count as much as anything in House races.

As the campaign enters its final days, we can develop a good sense of where the campaigns think they stand. That's because campaigns have two scarce resources. The first is money. We can't track money directly, but we can track what it is buying - namely, television advertisements. Thanks to Nielson, we know that McCain-Palin has recently focused its ad buys on Florida, Ohio, Pennsylvania, and Virginia. Meanwhile, its ad buys are being scaled back in Colorado.

The second scarce resource is time. If we follow where the candidates are traveling, we can know where they think their campaign events - and all the local media they generate - can make the best difference. So, let's examine where the McCain-Palin campaign has been travelling in the last few days. I built the following picture via Google Maps. It shows where John McCain has been from October 25th through today.

First off, we notice very few appearances out west. None in Nevada, none in Colorado, just two in New Mexico that occurred last Friday. Interestingly, there has been one trip to Iowa, which is curious, given the public polling. We also find three trips to Florida and one to North Carolina, which are consistent with his advertising in those states.

McCain's emphasis has been on Ohio. He's basically planted himself in the Buckeye State, holding eleven events in the last week.

Now, let's examine Sarah Palin's travels in the last week.

Palin has actually made no appearances in the west, another indication of the campaign's attitude about Colorado, Nevada, and New Mexico. Meanwhile, she has made two appearances in Iowa. Given that Obama is in the Hawkeye State today, we can infer that the race might be tightening there.

Palin, like McCain, has checked in with North Carolina and Florida. She's also been in Virginia and Indiana this week. However, Palin's emphasis has been on Pennsylvania.

So, the ticket has been focused on Ohio and Pennsylvania in the last week. Let's take a closer look at their travels in those two states. The following picture, again courtesy of Google Maps, does that. McCain visits are shaded blue. Palin's are green. Joint appearances are in purple.

Let's begin with a close look at Ohio.

What is interesting is that some of McCain-Palin's time has been spent in Democratic strongholds. For sure, McCain has made appearances in Defiance, a GOP stronghold in the northwest, and Lancaster, a GOP bastion in metro Columbus. However, there have been visits to blue-collar Democratic areas like Toledo, Elyria, and Steubenville. These are the kinds of places John Kerry was visiting at this point in the cycle in 2004. McCain-Palin has also hit several swing areas - Mentor (in metro Cleveland), Dayton, and Hanoverton (just south of Canton on I-77).

Overall, this has been a fairly balanced schedule, with both McCain and Palin toggling between Democratic strongholds, Republican strongholds, and swing areas. The only peculiarity I'd note is the relative inattention paid to the southern third of the state. Palin made a few stops in metro Cincinnati the week before, but McCain has not been there in quite a while. That is a bit surprising. I'm also surprised that there has been just one visit to Ohio's sixth congressional district (Steubenville). The sixth was a swing district in 2004 that heavily favored Clinton over Obama in the primaries. I would expect that, given fourteen total visits to the state, more than one would be dedicated to the sixth district.

What about the Keystone State?

There were fewer visits in Pennsylvania this week, and Palin hosted most of them. She held a rally in York, an exurban GOP stronghold that will need to come in big for McCain-Palin. She also has spent a good bit of time in the "Middle T" of the state, with visits to College Park (Correction: University Park), Williamsport, and Shippensburg. These are "rallying the base" visits with small town voters who have been loyal to the GOP for more than a hundred years (though Penn State helps Democrats in College Park). Note the visit to Erie, PA - a Democratic town that went for Kerry in 2004. Palin also made a stop in Latrobe, on the outskirts of metro Pittsburgh in Westmoreland County. She was in Beaver County the week before - a working class county also in metro Pittsburgh that has been trending toward the GOP in recent cycles. McCain, for his part, made a trip to Pottsville, PA - which sits about halfway between Harrisburg and Wilkes-Barre along I-81. They held a joint appearance in Hershey, PA - just outside Harrisburg.

Interestingly, what we have not seen from McCain-Palin in the last week are any trips to metro Philadelphia. The last time McCain was in metro Philly was October 21st for a rally in Bensalem. Palin was last there with a visit to Blue Bell on October 14th. This is an indication that the GOP ticket is focusing on the central and western parts of the state rather than the southeast. I discussed the logic of this earlier in the week.

So, what's the bottom line? Again, time and money are scarce resources. Candidates allocate them according to what they believe is the best strategy to win 270 Electors. That McCain-Palin has essentially planted itself in Ohio and Pennsylvania is an indication that it thinks its time is of good use here. Combine that with its enhanced advertising buys in these two states, and we can conclude that Ohio and Pennsylvania are crucial components of its electoral strategy.

Another interesting data point from Nielson, which has been tracking campaign commercials.

This is their chart of daily ads run in Colorado:

As we can see, McCain has been pulling back in the Centennial State. On Tuesday the 21st, his campaign ran 135 ads in the markets that Nielson monitors. This Tuesday, he ran just 55. That's a 60% decrease.

Meanwhile, here is what Pennsylvania looks like.

Last Tuesday, McCain ran 249 ads in monitored markets in the Keystone State. This Tuesday, he ran 337 ads. That's a 35% increase. Meanwhile, McCain has upped his ad buys in Ohio and Virginia.

What the heck is McCain up to? I have an idea. This is the picture of the primary results for Obama versus Clinton in Appalachia (courtesy of Sean Oxendine).

Meanwhile, the candidates are heading back to Iowa. Sarah Palin had two events there on Saturday, and Barack Obama will be there tomorrow. What's up with that?

Nielson reports that the combined household rating for Obama's campaign commercial last night was 21.7, meaning that 21.7% of all households in the 56 markets Nielson monitors watched the program. National numbers are still to come.

Here are the breakdowns for Obama's ratings in swing state markets:

Unsurprisingly, Philadelphia is the #1 market list here. Although interestingly enough Baltimore was the #1 market nationwide.

Check out more details here.

Michael Barone published an excellent essay on Pennsylvania earlier this week. In it, he analyzed McCain's prospects in the Keystone State, and from there pivoted to a broader discussion about the governing coalition Obama hopes to form.

I cannot recommend it more highly. Here, I'd like to supplement it, hoping to specify McCain's strategy in the Keystone State and its likelihood of success.

I suspect that McCain is aiming for a voting coalition roughly similar to the one Hillary Clinton fashioned in the April, 22 primary.

As we can see, Obama performed well in the southeast, winning three of the five counties in Metropolitan Philadelphia. Meanwhile, Hillary dominated the rest of the state, winning the rural areas, metro Pittsburgh, and the smaller cities along the Northeast Extension of the Pennsylvania Turnpike - Allentown, Bethlehem, Wilkes-Barre, and Scranton.

Indeed, another look at the primary data indicates just how well Clinton performed in these areas.

Those are some very large margins of victory outside Philadelphia. Pittsburgh presents an interesting case. Thanks to strong African American support in the city and the neighborhoods in the east, Obama won 45% of the vote in Allegheny County. However, outside Allegheny County, he did quite poorly. He only cracked the 33% threshold in Butler County, which is an exurban county (Obama tended to do well in exurban counties nationwide in the primaries). In Fayette County, he barely managed to win one of five primary voters. As we can see from the map above, Obama's generally performed quite poorly throughout the western third of the state.

This data seems to indicate an opportunity for John McCain and Sarah Palin. If primary voting Democrats broadly preferred Clinton over Obama, it might be that Independents and persuadable Democrats can be brought to the GOP. But can they really? Can McCain make any progress in these areas? The fact that both sides have been stumping in the state recently suggests that the answer might be yes. In light of this, let's examine the following chart - which tracks the "Republican tilt" of these areas since 1964.

As we can see, the rural areas have consistently supported the Republican candidate at levels greater than his national average. Allentown and the cities along the Turnpike Extension have been pretty consistent over the last 40 years, displaying a Democratic tilt between 1 and 10%. Meanwhile, Pittsburgh is once again an interesting case. Contrary to conventional wisdom, Pittsburgh is not home to "Reagan Democrats." The steel industry took a huge hit in the early 1980s, and the town responded by favoring Mondale by nearly 13 points. George W. Bush, however, did very well here, performing better than any Republican since Richard Nixon in 1972. That had a lot to do with the region's cultural conservatism.

So, there is potential for McCain. Republicans tend to do all right in these ares, so McCain might have an opportunity here to do well. If he hopes to win the state, he would have ultimately have to pull more votes in these three places combined than Bush did in 2004.

However, this strategy faces a major obstacle: Philadelphia. Barone notes that the southeast, according to the most recent Survey USA poll, seems primed to favor Barack Obama by a margin greater than what John Kerry pulled. He offers a very cogent theory for why:

My hypothesis is that that is because places like the Philly suburbs are places where the recent decline in household wealth has been most conspicuous. Housing prices mean a lot more to you when your house started off at $400,000 and declined to $290,000 than they did when you started off (as may be typical of Scranton or a blue-collar town in metro Pittsburgh) at $140,000 and declined to $110,000. Newspaper coverage of our current economic distress focuses on the very poor (like a recent Washington Post story on North Carolina, which focused on an ex-convict in a cheap motel in Charlotte), but the people who are getting hurt most visibly in their lifelong project of accumulating wealth are the more affluent. They're the ones whose house values have most visibly and spectacularly declined, and whose 401(k) accounts and stock portfolios have tanked in the last few months as well. Folks in Scranton or in the cheap motel in Charlotte didn't expect to live comfortably ever after off their increased house values, 401(k)'s, and Merrill Lynch accounts; a $700 monthly check from Social Security is about what they have long expected and that's not in danger (yet). Folks in the Philly suburbs did expect to live comfortably off such assets.

It seems to me that this hypothesis might be generalizable to McCain's weaknesses in Colorado, Virginia, and North Carolina - where wealthier suburbanites have taken a huge hit in the last month.

Barone notes another problem for McCain in Philly - as the above chart makes clear, the trend-line has been bad for the GOP. George H.W. Bush did quite well in metro Philly - but since then it has been downhill for the Republicans. George W. Bush lost the state twice in part because he twice lost 4 of the 5 counties in metro Philly. Additionally, as the primary data indicates, Obama did quite well here. He carried Philadelphia County thanks to strong support from African Americans. He also won Chester and Delaware counties, and he split Bucks County with Hillary Clinton. Clearly, the problems Obama had in the northeast and the southwest were not present in the southeast.

There might be an additional complication for McCain - the fast-growing counties surrounding metro Philly. Since 2000, the Census Bureau estimates that Berks, Lancaster, and York counties have grown anywhere between 5% and 10%. Most recent estimates indicate that more than 1.2 million people live in these places. Bush carried Berks County by about 6 points in 2004, and he dominated Kerry in Lancaster and York - winning both by near 2:1 margins. So, for McCain, victories in these places are already built into a statewide Republican loss. He must hold the line here. This could be problematic if Barone's hypothesis about the political implications of the financial crash is accurate.

So, all in all, Pennsylvania is a very tough challenge for McCain-Palin. Barone is skeptical that the GOP can win the state, and I share that skepticism.

That being said, I will be interested in examining the data after the election. I suspect the emphasis on Pennsylvania recently is due to the fact Obama is weak somewhere - either in the northeast, the southwest, or both. If that turns out to be the case on Election Day, Pennsylvania might exhibit a sharper urban-rural divide than it has in any recent cycle. As Barone notes - that would make for a very interesting electoral coalition for Obama, one based upon the wealthy and the poor in big cities like Philadelphia.

On Friday, I noted that the differences among the national polls is large enough to suspect that something other than random variation is causing the disagreements.

I'd like to expand on this point by examining today's Pew poll, which pegs McCain's share of the vote at 38%, with a margin of error of 3.5%. That means that Pew predicts with 95% confidence that McCain's true share of the vote is somewhere between 34.5% and 41.5%.

While we don't know McCain's true share of the vote, we do have an estimate of it - the RCP average. Right now, it puts McCain at 43.6%. This figure is far outside Pew's 95% confidence range. So, if we use the RCP average as our estimate of McCain's true share of the vote, we would conclude that Pew is an outlier.

The question then becomes whether it is outlying due to random variation, or some non-random cause. We can never know for sure, but we can make a few points.

First, the level of disagreement between the Pew poll and the RCP average is great. Indeed, if we assume that the Pew poll has an accurate read on the electorate, the chance that McCain's true share of the vote is 43.6% is less than 0.5%. Given the number of polls that cycle in and out of the RCP average, we should expect at least a few outliers. However, it would be pretty rare to find one that disagrees with the RCP average by such a large amount.

Second, the previous Pew poll, which had McCain at 39% of the vote, was also an outlier when compared against the RCP average. So, Pew has twice in a row pegged McCain's number at significantly less than the RCP average. It is very unlikely to see this kind of result if random variation is the only cause.

Does this mean that Pew is wrong? No. We could only conclude that Pew is wrong if we know McCain's true share of the vote right now. We don't know that. Instead, what we can conclude is that the difference between Pew and the RCP average is likely produced by something other than random variation.

Pew is not the only poll behaving in this fashion. Today, the Gallup traditional model pegs McCain's number significantly higher than the RCP average. It has done this several times over the last three weeks - and every day since it began it has shown McCain doing better than the RCP average. It is unlikely that random variation would produce these effects. Today's Rasmussen poll shows McCain significantly higher than the RCP average, and it has consistently been higher than the RCP average for the last three weeks. IBD/TIPP frequently pegs Obama's number significantly lower than the RCP average, and it has shown him lower than the RCP average every day since it began. The GWU/Battleground poll has shown McCain consistently higher than the RCP average for 10 of the last 10 release dates, frequently at significant levels.

None of this is consistent with what we would expect from random statistical variation. These considerations reinforce the point I made on Friday. In all likelihood, something else is going on here. The pollsters have different "visions" of what the electorate is, and these visions are inducing such divergent results.

This is why I would urge caution when interpreting all this polling data. We're talking about disagreements among good pollsters. I take all of these firms seriously whenever they produce new numbers. They are disagreeing with one another in ways that can't be chalked up to statistical "noise." That gives me great pause.

I've received several emails from people asking about the polls. The national polls do seem pretty variable, so I thought I would toss in my two cents on them.

First, we need a short primer on basic statistics. Real Clear Politics offers an unweighted average, or mean, of the polls. As long as there is more than one poll in the average, we can also calculate the standard deviation, which is one of the most important concepts in inferential statistics. The standard deviation simply tells us how much the polls are disagreeing with one another.

For instance, suppose we are testing the strength of Candidate A. We have 32 polls, which we can arrange graphically in what is called a histogram. Our horizontal axis shows the electoral strength of Candidate A. Our vertical axis shows how many polls we found with Candidate A pulling in that much of the vote. Let's say it looks like this:

The average is 50%. The standard deviation is 1.6, which basically implies that the typical distance between a given poll and the average is 1.6. That's a pretty small number, and it squares with how concentrated the polls are around this average.

Now, suppose we have a distribution that looks like this.

We get the same average, 50%. However, this time the observations are more dispersed around it. Here, the standard deviation is 3.0. That's higher, and we can see why. The individual polls vary more with one another. That's what the standard deviation shows us - how much the polls vary around the average.

A final point to get us ready. We might examine the spread between the two candidates - Obama is up 7 versus up 1 or what have you. This is certainly a valuable number to look at. Indeed, that's what we all care about! However, I am going to look at a candidate's share of the vote - not the spread. Ultimately, our analysis is going to rely upon each poll's reported margin of error. Those numbers do not refer to the spreads, but to each candidate's individual numbers. So, horse race polls actually have two margins of error - one for each candidate. Because the spread is the difference between them, it will be more variable than either candidate's individual numbers.

With this stuff in mind, let's focus on some hard numbers. As of this writing, Barack Obama's share of the vote in the RCP average is 50.3%. His standard deviation is 2.7. For McCain, whose average is 42.5%, the standard deviation is 2.3. For comparative purposes, I looked at the polls RCP was using from its 2004 averages. For roughly the same time in that cycle (10/17/04 to 10/24/04) Bush's standard deviation was 1.8; Kerry's was 1.7. This means that there is more disagreement among pollsters now than there was in 2004.

We can push this analysis further if we examine the distribution of each candidate's poll position. We'll first create a histogram of Obama's polling.

As we can see, most of the values cluster around the 49-54 range. However, there is a "tail" on the left-hand side. That's called a negative skew. That's a bit surprising. It's different from what we had in our stylized pictures for Candidate A.

Now, let's examine distribution of McCain's support.

There's no tail here, but the picture is still somewhat surprising. They are spread out fairly evenly across a broad range of values, with little clustering in the center.

Of course, a visual inspection can only take us so far. When we have only a few observations - and here we "only" have 15 - the true shape of the picture might not be clear. If we were to add another 5 or so polls, we might see something more like those stylized pictures presented above.

So, let's push the analysis a little bit further by looking at specific polls. We can test to see if the polls are separated from the average by a statistically significant amount. Again, since we're dealing with each candidate's individual poll positions - we'll test each candidate's number in an individual poll against the RCP average. To make sure we dot all our "i's" and cross all our "t's," we'll supplement the RCP average with a weighted average of the polls, which takes into account the number of observations when averaging the polls together.

Of the fifteen polls in the RCP average, four fall significantly outside the average for Obama and five do so for McCain. Meanwhile, three polls are right at the boundary of significance (one for Obama, two for McCain). The rules of statistics being what they are, we should expect a few polls here or there to fall outside the average by a statistically significant amount. But this is a lot. 40% of all our tests produced results around or outside the acceptable range.

So, we have made three observations: (a) relative to 2004, the standard deviation for Obama and McCain's polls are high, indicating more disagreement among pollsters at a similar point in this cycle; (b) the shape of the distribution of each candidate's poll position is not what we might expect; (c) multiple polls are separated from the RCP average by statistically significant differences.

Combined, these considerations suggest that this variation cannot be chalked up to typical statistical "noise." Instead, it is more likely that pollsters are disagreeing with each other in their sampling methodologies. In other words, different pollsters have different "visions" of what the electorate will look like on November 4th, and these visions are affecting their results.

Think of it this way. Suppose there is a bag of 130 million red and blue marbles that all the pollsters are sampling from. One pollster will pull a sample of 750 marbles, another a sample of 2,500, and so on. Oftentimes, they are going to pull different results from the bag. One pollster might pull 53% blue, another might pull 52%, and so on. However, as long as they are all pulling marbles from the same bag, the results will probably not differ too wildly. And after enough time, the distribution of those pulls should look something like those idealized pictures of Candidate A.

However, what if each pollster had a slightly different bag s/he was pulling from? In that situation, we should find more divergent results. That's basically what I'm suggesting here - that the bags the pollsters are pulling from are different. That's producing some of these larger-than-expected variations.

Now, I want to be clear: I am not making any claims about which pollster has the better sample of the electorate. I'm not singling anybody out for being right or wrong because frankly I do not know. I'm just pointing out that there seems to be disagreements among them that cannot be explained by random variation.

Importantly, there is one thing that the polls do not disagree on, the fact that Obama has a lead. All the polls show that. Also, we might begin to see convergence here soon. If pollsters have different methods for predicting what the electorate will look like, those methods might produce similar-looking "electorates" by the time we get to Election Day. At least for now, though, there is disagreement - not about who has the lead, but about how big that lead is.

The current Real Clear Politics national average gives Barack Obama a 5.2-point lead over John McCain. This makes a comeback for McCain quite difficult, but not inconceivable.

The election of 2000 provides a good reminder of this fact. Like 2008, that was a year in which there was no incumbent on the ballot. Political observers were treated to some wild swings in the polling, right up to the very end. The following graph makes this clear by charting a daily average of nonpartisan polls up to November 6th. It also includes the final results from Election Day.

By this metric, Gore was down by 5.5 points on October 21st. Even as late as November 5th, he trailed by 4.7 points. However, Gore finished well. Though he was as low as 42% in late October, he pulled in an extra 6.5% of the vote in the last week to sprint ahead of Bush by Election Day. Meanwhile, Bush's final share of the vote was only a point or so higher than what he polled in the final week of the campaign.

We should not expect the 2008 election to evolve like the 2000 contest. Presidential elections are all unique, and each plays out in its own peculiar way. For starters, that DUI story was certainly no help to Bush. Another big difference between 2008 and 2000 is the Nader factor. Nader was polling about 4% of the vote prior to Election Day, and he ultimately won 2.7%. That's an indication that, at the last minute, some of the Nader voters switched to Gore, thus enhancing his rally. So, to win the popular vote, McCain would have to surge better than Gore did. That would be no small feat.

Instead of being a strict parallel, the election of 2000 offers two lessons. First, October can be a volatile month, as up to 30% of the electorate is making its final choice. Where you are on October 20th could be quite different than where you are on October 31st. Second, Obama must focus on "closing the deal." That is, he must keep his soft supporters on his side, and convince enough of the remaining undecideds to go for him. The above graph shows that Bush failed to do this eight years ago. What appeared to be a solid win for the Texas governor ended up as a kind of technical victory.

Obama had a very hard time closing the deal in the primaries. Fresh off his win in the Iowa caucus, he headed into the New Hampshire primary with an 8-point lead in the RCP average. Yet Hillary Clinton won by 3 points. Either all the polls were wrong, or there was a big shift to Clinton just before the primary. My bet is the latter.

Obama, of course, rebounded from his New Hampshire loss - and held his own on Super Tuesday. He then went on a fantastic run for the rest of February, winning the caucuses in Louisiana, Nebraska, Washington, Maine, and Hawaii. He crushed Clinton in the DC, Maryland, and Virginia primaries on February 12th. He followed that up with a big win in Wisconsin a week later. Clinton's back was up against the wall heading into the Texas and Ohio primaries on March 4th. It seemed like it was over. Undecided Democrats would see Obama as the one bound to win, and they'd rally behind him. In a word, it seemed like Obama would finally pick up some momentum.

It didn't play out that way. Clinton held Texas and won a decisive victory in Ohio. Despite Obama's delegate lead, and word from pundits everywhere that the race was over, he never developed any momentum. Even after his surprisingly large win in North Carolina on May 6th, Clinton won decisive victories in West Virginia, Kentucky, and Puerto Rico. On June 3rd, when everybody knew it was over, South Dakota favored Clinton over Obama by 10 points.

In fact, during the three final months of the primaries, Hillary beat him 52-48. Obama never scored a knockout against her. Instead, he won the nomination by lining up the party establishment and securing oversized victories in the caucuses, for which the Clinton campaign was unprepared.

Obama's weak finish in the primaries serves as the subtext for these next two weeks. Republicans are wondering if voters will again have second thoughts about him now that he's the clear frontrunner. Meanwhile, the Obama-Biden campaign is clearly taking no chances. Here in Western Pennsylvania, as in most swing states, Obama is virtually omnipresent on television. That's the mark of a campaign that learned a tough lesson in the primaries: there will be no coasting to victory.

Many of the Obama advertisements running here are actually autobiographical, taking snippets from his convention biographical video and repackaging them in 30-second spots, with a distinct emphasis on his American roots. That's not what we should expect so late in the cycle - but Obama is no typical candidate. The point of such ads is to make voters feel comfortable with the idea of him as President. I think this ultimately kept him from landing that knockout blow against Hillary. Those late-breaking primary voters felt more at ease with Hillary instead of Obama. That's what Obama wants to avoid next month.

Obama's advertising edge might be moving numbers in individual states, but his national number has stayed pretty constant in the last few weeks. Here's a track of the RCP average from October 1st through today.

As we can see, Obama's numbers have been been stable between 48.8% and 50.2% this month, with an average of 49.4%. McCain's numbers have been more variable, ranging between 42% and 44.3%. Obama jumped out to a 6-, 7-, even 8-point lead not because his support base was growing. Instead, McCain's was shrinking - though it appears to be coming back a bit.

Meanwhile, the undecided share of the electorate seems to have been pretty constant at about 7% of all voters. There are plenty of reasons to expect these people to decide late, regardless of how intense the electioneering becomes. So, even if Obama's ad blitz has not moved these people yet - it could eventually move them. And again, it might be moving undecideds in some states.

If Obama's job between now and November 4th is to make these remaining undecideds comfortable with the idea of him as President - McCain must make them feel uncomfortable, casting himself as a safe alternative. How will he do that? It won't be via Jeremiah Wright. McCain has ruled that out. [Update: Maybe not?] He might continue to push William Ayers or others from Obama's Chicago past, but I doubt they will be front-and-center. We're hearing the "s word" - socialist - from McCain surrogates these days, and that will probably continue to be a refrain his campaign sings.

My hunch is that McCain's closing argument will focus on Obama's relative inexperience. There are three reasons to think this. First, this attack seemed to work for Hillary in late February. When Obama was poised to win the nomination, she released that "3 AM" ad, which many took to be quite effective. Second, unlike Wright and Ayers, it is not in the least bit radioactive, so there is little risk of backlash. Third, it's Obama's principal weakness. By conventional metrics, Obama is to be counted as one of the least "credentialed" nominees we have had in a long time, ranking with Thomas Dewey and William Jennings Bryan.

Remember this web ad? McCain dropped it during Obama's convention, but then it disappeared from view. My guess is we'll see more stuff like this before November 4th - especially the last segment where Obama talks about whether he would run in 2008.

Last night's debate was a great demonstration of how different the political styles of these two candidates are. I'd wager that one's opinion on who "won" depends on the style one prefers.

John McCain comes across as a direct man who is inclined to speak the thoughts that come into his head. He relies upon short, blunt sentences unadorned with adjectives and adverbs. This makes him seem like a man who lacks guile. This is a good quality for a politician to have, and that showed through last night.

But this style carries with it some problems. His answers to questions last night had a tendency to be unconnected from one another (as they have been generally in these debates). He's wont to toss in a reference to energy independence when the discussion is health care or housing or whatever. All in all, his responses tend not to hang together terribly well - thus diminishing their forcefulness.

Another problem for McCain is that he often references things that might go over people's heads. For instance, when talking about killing defense spending, he'll reference "DOD." That's a bit of an insider term, something that he's familiar with, but something that swing voters might not get. Similarly, he'll often talk about "reauthorization" and so on. He's been in the Senate for a long time, and he's picked up the language of its culture. This also diminishes his rhetorical thrust. Several times last night, notably during the abortion debate portion, McCain offered a retort to Obama, but he couched it in technical language, I was left thinking that it didn't work.

Barack Obama's style is quite different. He speaks in paragraph form. He gives lists and categories when answering his questions, and he's frequently inclined to place a specific subject in a broader context. More than anything, this has helped him ameliorate concerns that he's unprepared. This speaking style makes him seem like he knows what he is talking about.

However, his speaking style makes him seem professorial. I'd contrast him with Bill Clinton, who also relied on paragraphs, but who used them as a way to "talk." Obama doesn't talk so much as he "lectures." For sure, he's more "talkative" now than he was during the primaries - he had quite a number of good, foksy lines last night - but he still sounds like he's giving a college lecture.

He also sounds a little lawyerly. His paragraphs often contain "fine print" that needs to be considered. That means that, for as knowledgeable as he sounds, its difficult to identify his precise position. For instance, last night he spoke about how he would involve the federal government in local schools - but he introduced those comments by praising local control. It leaves one wondering exactly what his vision for public education is.

One thing that these two men seem to have in common: they don't like each other so much. This was a striking episode from last night:

Scheiffer: All right. Would you like to ask [Obama] a question?

McCain: No. I would like to mention that...

Obama gave as he good as he got on that front. Several times he chuckled at some point McCain was making. I noted this in particular during the back-and-forth about who's running the dirtier campaign. McCain closed one response by saying that his campaign is "about getting this economy back on track." Obama chuckled at that.

I think that neither man helped themselves with such behavior.

A reader writes in with a question:

A few of my friends have said almost in the same words, "I was thinking of voting for McCain until he chose Palin. After that, I'm voting for Obama." Packer's article in the New Yorker on Ohio also quotes someone saying the same thing.

So I was wondering if you have any insight into the Palin Effect. Any polls out that are getting to this issue somehow?

This is a good question.

One thing we can say is that the electorate's reaction to Palin has been polarized. Rasmussen finds that 35% of voters have a very favorable view of Palin, compared to 33% with a very unfavorable view. For Biden, those numbers are 25% and 21%, respectively. LA Times/Bloomberg finds that Republicans overwhelmingly like her, Democrats overwhelmingly don't, and Independents are split evenly, 41-41. This ambivalence among Independents might be due to perceptions that she is not prepared to take on the responsibilities of the presidency. The Diageo/Hotline poll finds that by a 44-53 margin, voters don't think she'd be ready for that role.

LA Times/Bloomberg also finds the following:

Last month's poll showed that slightly more voters were more likely to vote for McCain because of Sarah Palin's presence on the ticket. This poll shows that the Palin rock star status has waned (except with the Republican's core base). More voters are less likely to vote for McCain because of her presence on the ticket. Independents have flipped their allegiance. In September 38% said they were more likely to vote for McCain because of Palin - now just 19% say that. Women were split last month, but now lean toward less likely, as do men.

The polls show movement consistent with what the reader hypothesizes, but we have to be careful. Media polls have a habit of oversimplifying public opinion - and the write-ups of the news outlets who commission them only exacerbate this tendency. Frankly, I don't think these "more likely, less likely, no difference" questions have much value. The reason is that it matters who you are likely to support to begin with. Suppose you're a solid Obama supporter, and you say Palin makes you less likely to vote for McCain. That isn't really accurate. You're likelihood of voting McCain was already 0%. You can't go any lower. So it goes for McCain supporters, too. That means that these results don't reflect voter sentiment on the precise question, but something slightly different.

So, if enhanced anxiety about the economy has pushed you from leaning McCain to undecided, or from undecided to leaning Obama, how will you answer this question about Palin? If you're feeling more negative about the McCain-Palin ticket these days, you might say "less likely" even if she is only an ancillary factor in your decision-making. If that's the case, and the analyst is not careful, he will draw your causal arrow in the wrong direction. It's not that your aversion to Palin makes you less likely to vote for McCain. It's that you're less likely to vote for McCain so you're more averse to Palin.

I'm not saying this is what's really going on. I think this is a reasonable alternative hypothesis to the reader's "Palin Effect" theory. Complicating matters even more is the fact that the two ideas are not mutually exclusive. Some voters might be moving from McCain because of Palin. Other voters might be moving away from McCain because of the economy, and that's eroding Palin's numbers. There might be other explanations, too - so overall we have a tricky situation on our hands.

With that in mind, let's look at the LA Times/Bloomberg breakdown:

The LA Times was right to note that Palin's "less likely" number among Independents has increased by 13%, but they fail to note that the overall number of Independents who are unmoved either way has increased by 9%. Democrats also see her as less relevant to their vote choices, while she has become slightly more so to Republicans.

This might be a problem for the Palin Effect theory. If she's a major factor in McCain's drop-off, why is she becoming less salient? It seems to me that if she's pushing some voters away from McCain, she'd be more relevant to the thinking of voters generally, and we'd see a decrease in the number who say "No Difference."

Another issue - in most cycles, voters don't make a choice based on the veep selection. The veeps have a big, splashy rollout - but afterwards they become part of the backdrop. Should we really expect, with everything else that is happening this cycle, Palin's effect will be substantially different?

Unfortunately, media polling does not probe deep enough to give us final answers. So, I'll offer my guess about what is going on. I think there might be a Palin Effect among some Independent voters, consistent with what the emailer is finding among his friends. Some group of Independents might still be with McCain had he picked Pawlenty, but Palin has turned them off.

However, I'd wager this effect is pretty marginal. Above all, I think there is something bigger going on this cycle, influencing voters in ways that the personalities and events of the day don't. This is the alternative hypothesis I offered above. A rising tide lifts all boats, but a falling tide lowers them - and Palin's numbers have slid as the GOP ticket has run into bigger trouble. I'd note that LA Times/Bloomberg shows McCain has suffered a broad erosion in his support among Independents. In September, McCain led Independents 49-34. Today Obama leads 44-39. That's a 20-point swing, a sign that something more dramatic is happening in this race.

I'd note also that when we talk about a Palin Effect, we should remember that what matters is the net effect. Even if Palin has hurt McCain a bit among Independents, I think she is keeping morale up among Republicans. In the LA Times/Bloomberg poll, 49% of Republicans had a very positive view of McCain while 35% had a somewhat positive view of him. For Palin, 61% of Republicans had a very positive view compared to 23% with a somewhat positive view. This squares with the evidence we've seen from the McCain-Palin rallies. It also squares with the RNC's fundraising in the wake of the Palin pick. These are benefits to be counted against any Independent voters who have peeled away from McCain because of her.

Given the situation the GOP currently finds itself in, the fact that she rallies the base might help the party on Election Day. McCain's poll position has slipped, and most people now believe that Obama will win. If this belief persists through Election Day, Republicans might be less likely to come to the polls, thereby damaging down-ballot candidates. If Sarah Palin can give Republicans a reason to come out and vote, that might make her presence a net benefit even if she is driving away a few Independents from the GOP ticket.